Muon Tracking and Calibration in PHENIX

In order to make sense of the data you need ...

This research focuses on the development of software for the muon tracking and calibration
subsystems of the PHENIX detector at RHIC. The purpose of the muon arms is to measure both
the mass and momentum of vector mesons and dimuon pairs to measure the properties of
quark-gluon plasma. Each muon arm consists of three stations of tracking chambers mounted
inside an end-cap magnet followed by a muon identifier system which consists of six layers of
steel absorber interspersed with tracking detectors. Few particles pass through the absorbing
plates, at which point certain non-showering particles at consistant depths are identified as
muons. Then pattern-recognition algorithms reconstruct the particle's expanding helical track
from the thousands of hits recorded at the cathode strip stations. Additionally, the relative
ADC calibration of cathode strips within a chamber must be known to within 1% to obtain a
100 micrometer resolution from the cathode planes. To achieve this, all 16,000 channels of
cathode strips must be pulsed at roughly 30 different voltage levels to determine the gain
calibration, while the pedestal calibration requires data points taken for each strip at zero
voltage. Finally, these calibration data are analyzed and the approximately 80,000 calibration
constants are stored in a database.

1)MuTr Cathode Strip Calibration System Information

2) MuTr calibration Database & Objectivity : local setup in Yonsei

3) MuTr calibration Database & Objectivity : setup and test at phnxmut.rch.rhic.bnl.gov
Source Code

Calibration system for PHENIX muTr cathode strips

calibration parameters.

       /gain - gain files,
        /pedestal - pedestal files,
        /bad_channel - bad channel files, and
        /map - some special files.

Refreences

ADC values larger than 250 [FS=1023] are not included in any pedestal calculation

Gain determination

TPC group will provide a pad gain correction table Gaux[384], one value per pad on a mezzanine. This table will be used to correct the empirically
determined gain (see below) to correct for end pad problems. This gain correction has a required dynamic range of <2:1 [Initial table will be all
unity gain factors.]

A gain determination run will consist of approx 100 pulses with the same nominal amplitude.

Run Control will supply a window of timebin numbers to be used in the selection of the true peak from the pulser. The sequence used will be the
sequence within the given timebin window which has the largest single ADC value and which contains at least 3 timebins above threshold.

All operations on the ADC value described will be based on the expanded ADC scale (8 bit values will be restored to 10 bits).

For each pad and for each pulse, the selected sequence will be used to determine ADC-weighted mean timebin number t0 = S ti x ADCi
/SADCi,
and sum of ADC values SADC.

At the end of the run, the ensemble quantities

        <t0>   average t0 for all events,
        (t0) rms rms for t0 over all events,
        <SADC> sum of ADC values averaged over all events,
        (SADC) rms rms for ADC sum over all events

are calculated. A relative gain is calculated for each pad:

Gi = <SADC>events / <SADC>pads,events

[ratio of ADC sum averaged over all events to ADC sum averaged over events as well as pads in this mezzanine board]. Any Gi significantly
different from the average over all 384 pads is deleted from the list and the average is calculated again. The recalculated mean is then used to
calculate a new set {Gi}. The multiplicative factor to be used to modify the translation table for a given pad is given by

(Gapplied)i = (G)i x (Gaux)i

where (Gaux)i is given by run control [see above].

Bad channel determination

A timebin will be considered bad if

    the pedestal RMS < 0.05 counts
    the pedestal RMS > 25 counts
    pedestal value is out of range [2 sigma from mean for the pad]

For timebins 20-400 if there are >50% bad timebins the channel is declared to be bad

Gain will not be used to determine that a channel is bad.

1 . Pedestal Corrections (http://pibeta.phys.virginia.edu/~pibeta/docs/publications/penny_diss/node33.html)